Do you remember the following symbols?
You might seen them in your school english books and Dictionary.
See here to remember few more symbols.
These are the symbolic representation of the sounds of any language.
Using these symbols, anyone can read the strings and spell like the original sound of the word, even though he dont know the language.
This IPA symbols are available for all languages, as they represent only the various sounds.
I am thinking of an open source system for building Text to Speech System for Tamil, ( actually for any language)
Mr. Vinoth created a web application in PHP, which can convert a Tamil string to IPA.
Here it is – http://anunaadam.appspot.com/
My brother Arulalan convert this as a Python Library.
Now, we need all the sound files for the Tamil IPA symbols.
Where can we get the IPA symbols for Tamil ?
You can click on the symbol on wikipedia to know how a symbol sounds.
Hear the audio file. It is an full example.
But, we can not use these audio files for Text to IPA, as they have very long extended audio.
Now, we need the short and clear audio files all the tamil symbols, just like how we spell the words and letters on our daily speech.
As we have text to IPA convertor, once we have all the sound files with regular sounds, we can play them together to create almost nearly speech version of any word.
This is my assumption only. There may be issues. But hope we can fix them all.
If we have multiple voices, then our Text to speech engine can have multiple voices too.
I request you to record the IPA sounds for Tamil symbols and send to me.
My email : tshrinivasan AT gmail DOT com
If this plan works well, we can extend this to all the languages.
Share your thoughts on this.